Collecting the evidence

Question 1

Constructing a question - PICO

Answer 1

Patient

Intervention Comparison Outcome

Question 2

‘Anatomy’ of a good question

Answer 2

1. Define precisely whom the question is about (how would I describe a similar group of patients)
2. Define which option you are considering (eg drug treatment) and possible comparison (eg placebo or standard therapy)
3. Define the desired (or undesired) outcome (eg reduced mortality, better quality of life)

Question 3

• Systematic Reviews/ Meta-analyses

Answer 3

– Secondary sources of information vetted by independent

researchers and clinicians (Cochrane Collaboration)

Question 4

• ClinicalPracticeGuidelines

Answer 4

– Reviews covering large disease groups and treatment

strategies (NICE guidelines, SIGN Guidelines)

Question 5

• Original article containing primary research data

Answer 5

– eg randomised-controlled trial (RCT)

Question 6

When trying to answer a clinical question

Answer 6

1. Start with
– Cochrane Reviews
– NICE and SIGN Guidelines
2. Then use
– MedLine (Ovid, PubMed)

Question 7

Interpreting the evidence

Answer 7

• How are clinical studies designed
• Different types of data
• Hypothesis testing
• P value and statistical significance
• Type I and Type II errors
• Confidence intervals
• Number needed to treat(NNT)
• Odds ratio
• Forest plots

Question 8

Hypothesis construction

Answer 8

– Null hypothesis– two sets of data are from the same population and not different

– Alternative hypothesis– two sets of data are from different populations and are different

Question 9

quantitive discrete

Answer 9

can only have certain numerical

values (number of children)

Question 10

quantitive continuous

Answer 10

do not have discrete steps (height and weight)

Question 11

Nominal (unordered categories)

Answer 11

• Male/female, green/blue/ eyes, alive or dead

Question 12

Ordinal (ordered categories)

Answer 12

• Objective – heavy, moderate or light drinkers (based on the number of units of alcohol drunk per week), grade of breast cancer

Question 13

questionnaires

Answer 13

• Subjective

Question 14

Hypothesis testing

Answer 14

– Assume the null hypothesis – two sets of data are from the same population and not different
– Determine the probability that the null hypothesis is correct – P value

Question 15

p value

Answer 15

if this is one it means the null hypothesis is true

Question 16

P < 0.05

Answer 16

• An arbitrary cut-off of 0.05 or 5% (1 chance in 20) has been chosen to indicate that the null hypothesis can be reasonably rejected
• If P < 0.05 then there is a statistically significant difference

Question 17

Type I error

Answer 17

rejecting the null hypothesis when it is true (false positive) - concluding there is an effect when there isn’t (P is small)

Question 18

Type II error

Answer 18

not rejecting the null hypothesis when it is false (false negative) - concluding there is no effect when there is (P is large)

Question 19

Power

Answer 19

• The power of a test is its ability to reject the null hypothesis when it is false
– the capacity to detect an effect if one is present

Question 20

power questions

Answer 20

– Was the sample size large enough? – Variation small enough?